In lieu of an abstract, here is a brief excerpt of the content:

C h a p t e r 9 Monitoring Urban Growth and Its Environmental Impacts Using Remote Sensing: Examples from China and India Karen C. Seto The size of the world’s growing urban population gives urgency to the need for accurate estimates of the location, size, and growth of existing urban areas as well as forecasts of likely regions, magnitudes, and con- figurations of future urban growth. However, to date, there exists no global database that accurately describes and maps which portions of Earth’s habitable land are urbanized, or how those portions have changed over the recent decades. Satellite remote sensing and spatial modeling offer tremendous opportunities to map historical patterns of urban growth, monitor urban areas, and forecast urban expansion. Satellite -based efforts at mapping global urban extents fail to agree on the size and pattern of urban land use, with estimates ranging from 0.2 percent to 2.4 percent of terrestrial land surface circa 2000 (Potere and Schneider 2007). Recent advances in remote sensing—both in satellite hardware technology and image processing algorithm development—provide opportunities for collection and dissemination of timely information on urban form and size that can be useful for policy and planning. In spite of these developments, there are also limitations to remote sensing and its application in practice. In this chapter, I will describe some of the opportunities for, and limitations on, monitoring urban growth using remote sensing data, and I will provide examples of the environmental impacts of urban growth, as monitored with remote sensing. Satellite Remote Sensing: Opportunities and Limitations for Urban Mapping Satellite remote sensing affords a number of unique opportunities for monitoring urban growth. The internally consistent measurements and 152 Urban Spatial Growth and Development long observational record of satellite sensor data make it an attractive source of reliable information on urban extent and form. Beginning with the launch of the first Landsat satellite in 1972 and continuing through Landsat 7, satellites have provided more than thirty years of 30–80 m multispectral imagery for much of Earth’s surface. Each Landsat scene covers approximately 170 km north-south and 185 km eastwest , an area that easily encompasses a metropolitan area if the city is imaged near the center of the scene. Satellite images are digital data of reflected energy collected across portions of the electromagnetic spectrum. Because most satellite data are multispectral, they contain information from the nonvisible portions of the electromagnetic spectrum (vegetation and soils are most reflective in the nonvisible range). Among the many types of information that can be derived from remote sensing, those relevant to the discussion of urbanization include local surface temperature, wildlife habitats and biodiversity corridors, and extent of impervious surfaces. Moreover, because satellite images are simply digital arrays of information, they can be reprocessed in the future as new digital image processing methods become available. The ability of satellite data to identify urban areas rests on the unique spectral characteristics of urban areas relative to other land covers such as vegetation, water, or soil. Because urban areas are composites of other land covers (e.g., lawns, swimming pools, rooftops, concrete sidewalks, buildings, etc.), a single ‘‘urban pixel’’ in an image is likely to be a mix of composite land covers. Very few urban pixels will be ‘‘pure’’ (i.e., entirely pavement, entirely building, entirely roads). The purity of any given pixel will be determined by the scale of the urban elements (e.g., building, road) relative to the spatial resolution in the image (Woodcock and Strahler 1987). For example, the spatial resolution of Landsat makes it useful for mapping large urban areas and indicators of urban form and land use, but it does not lend itself to street-level urban mapping or differentiating between residential and commercial urban development , or between high-density and low-density urban development, unless additional ancillary data are available. Data from the first Landsat period, from 1972 to 1983, were imaged at 68 by 83 meters and commonly resampled to 57-meter spatial resolution. Since 1984 and Landsat 5, multispectral data have been collected at 30-meter resolution. For the purposes of urban mapping, these data are relatively coarse and cannot detect small-scale urban change. For example, intercity highways, isolated patches of small urban development, or urban infilling may not be distinguishable in a Landsat image. Where urban growth is occurring in agricultural regions, Landsat data may be too coarse to differentiate Using Remote Sensing in China and India...

pdf

Back To Top

This website uses cookies to ensure you get the best experience on our website. Without cookies your experience may not be seamless.